# Session 1 Handout: Asking Well: Prompting Basics

<p class="meta">FAST-16 Working Well with AI · 7formations · Instructor: Wadood Chaudhary · All examples fictional · Suggested reading: Svekis, AI Literacy for All Employees (2026), chapter 6</p>

## Why this session matters

The answer you get from an AI assistant can only be as specific as the request you give it. A vague prompt gets a generic answer, and generic answers are where invented details creep in. Good prompting is not a trick; it is the same skill as giving a new coworker a clear assignment.

## Key concepts

### 1. Weak versus strong prompts

A weak prompt is short, vague, and leaves the AI to guess: "Write about parking." A strong prompt says who it is for, what to do, what facts to use, and what the result should look like.

### 2. The four elements

| Element | Question it answers | Cape Harbor Inn example (fictional) |
| --- | --- | --- |
| Context | What is the situation? What facts apply? | "Overflow parking moves to the town lot on Main Street from June 1, a 4-minute walk, free with a desk pass." |
| Specific instruction | What exactly should the AI do? | "Write a short FAQ answer to 'Where do I park?'" |
| Structure | What form, length, format? | "Three sentences, under 70 words." |
| Audience | Who will read it? | "Arriving guests who have never visited." |

Add a **gap rule**: "Use only these facts; if something is missing, write [CHECK]." It is the single most useful line for cutting invented details.

### 3. Four techniques

- **Step-by-step:** break the task into ordered steps.
- **Iterative:** improve with specific follow-ups ("keep the facts, cut to two sentences").
- **Role:** "as a friendly front-desk lead" changes style and focus, not expertise. You are still accountable.
- **Constraint:** limits on length, scope, source, or tone.

### 4. Common mistakes

Expecting one vague prompt to be perfect; leaving out the facts; asking for several unrelated things at once; accepting the first answer; and keeping one long chat for every task. Start a **new chat** when the task changes or when the thread contains something that should not be there (and report the slip).

## Worked example

Weak: "Reply to a guest who wants to bring pets." Typical result: a warm reply that invents a $50 fee and welcomes cats. Strong: "Using only the pet policy below, write a friendly three-sentence reply to a guest asking to bring two dogs and a cat. If the policy doesn't cover something, say we'll check with the manager. --- Dogs up to 40 lb, garden rooms only, $35 per night; no cats. ---" Result: accurate, polite, and nothing invented.

## Try this before session 2

- Rewrite one prompt you used recently with all four elements and a gap rule.
- Read Svekis chapter 6.